* [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel
@ 2021-04-22 12:34 Niklas Söderlund
2021-06-10 12:45 ` Daniel Lezcano
2021-08-26 16:25 ` [tip: timers/core] clocksource/drivers/sh_cmt: " tip-bot2 for Phong Hoang
0 siblings, 2 replies; 5+ messages in thread
From: Niklas Söderlund @ 2021-04-22 12:34 UTC (permalink / raw)
To: Daniel Lezcano, Wolfram Sang, linux-kernel
Cc: linux-renesas-soc, Phong Hoang, Niklas Söderlund
From: Phong Hoang <phong.hoang.wz@renesas.com>
If CMT instance has at least two channels, one channel will be used
as a clock source and another one used as a clock event device.
In that case, IRQ is not requested for clock source channel so
sh_cmt_clock_event_program_verify() might work incorrectly.
Besides, when a channel is only used for clock source, don't need to
re-set the next match_value since it should be maximum timeout as
it still is.
On the other hand, due to no IRQ, total_cycles is not counted up
when reaches compare match time (timer counter resets to zero),
so sh_cmt_clocksource_read() returns unexpected value.
Therefore, use 64-bit clocksoure's mask for 32-bit or 16-bit variants
will also lead to wrong delta calculation. Hence, this mask should
correspond to timer counter width, and above function just returns
the raw value of timer counter register.
Fixes: bfa76bb12f23 ("clocksource: sh_cmt: Request IRQ for clock event device only")
Fixes: 37e7742c55ba ("clocksource/drivers/sh_cmt: Fix clocksource width for 32-bit machines")
Signed-off-by: Phong Hoang <phong.hoang.wz@renesas.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
---
drivers/clocksource/sh_cmt.c | 30 ++++++++++++++++++------------
1 file changed, 18 insertions(+), 12 deletions(-)
diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c
index c98f8851fd680454..cadd09ad1a0946b9 100644
--- a/drivers/clocksource/sh_cmt.c
+++ b/drivers/clocksource/sh_cmt.c
@@ -578,7 +578,8 @@ static int sh_cmt_start(struct sh_cmt_channel *ch, unsigned long flag)
ch->flags |= flag;
/* setup timeout if no clockevent */
- if ((flag == FLAG_CLOCKSOURCE) && (!(ch->flags & FLAG_CLOCKEVENT)))
+ if (ch->cmt->num_channels == 1 &&
+ flag == FLAG_CLOCKSOURCE && (!(ch->flags & FLAG_CLOCKEVENT)))
__sh_cmt_set_next(ch, ch->max_match_value);
out:
raw_spin_unlock_irqrestore(&ch->lock, flags);
@@ -620,20 +621,25 @@ static struct sh_cmt_channel *cs_to_sh_cmt(struct clocksource *cs)
static u64 sh_cmt_clocksource_read(struct clocksource *cs)
{
struct sh_cmt_channel *ch = cs_to_sh_cmt(cs);
- unsigned long flags;
u32 has_wrapped;
- u64 value;
- u32 raw;
- raw_spin_lock_irqsave(&ch->lock, flags);
- value = ch->total_cycles;
- raw = sh_cmt_get_counter(ch, &has_wrapped);
+ if (ch->cmt->num_channels == 1) {
+ unsigned long flags;
+ u64 value;
+ u32 raw;
- if (unlikely(has_wrapped))
- raw += ch->match_value + 1;
- raw_spin_unlock_irqrestore(&ch->lock, flags);
+ raw_spin_lock_irqsave(&ch->lock, flags);
+ value = ch->total_cycles;
+ raw = sh_cmt_get_counter(ch, &has_wrapped);
- return value + raw;
+ if (unlikely(has_wrapped))
+ raw += ch->match_value + 1;
+ raw_spin_unlock_irqrestore(&ch->lock, flags);
+
+ return value + raw;
+ }
+
+ return sh_cmt_get_counter(ch, &has_wrapped);
}
static int sh_cmt_clocksource_enable(struct clocksource *cs)
@@ -696,7 +702,7 @@ static int sh_cmt_register_clocksource(struct sh_cmt_channel *ch,
cs->disable = sh_cmt_clocksource_disable;
cs->suspend = sh_cmt_clocksource_suspend;
cs->resume = sh_cmt_clocksource_resume;
- cs->mask = CLOCKSOURCE_MASK(sizeof(u64) * 8);
+ cs->mask = CLOCKSOURCE_MASK(ch->cmt->info->width);
cs->flags = CLOCK_SOURCE_IS_CONTINUOUS;
dev_info(&ch->cmt->pdev->dev, "ch%u: used as clock source\n",
--
2.31.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel
2021-04-22 12:34 [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel Niklas Söderlund
@ 2021-06-10 12:45 ` Daniel Lezcano
2021-06-11 7:56 ` Niklas Söderlund
2021-08-26 16:25 ` [tip: timers/core] clocksource/drivers/sh_cmt: " tip-bot2 for Phong Hoang
1 sibling, 1 reply; 5+ messages in thread
From: Daniel Lezcano @ 2021-06-10 12:45 UTC (permalink / raw)
To: Niklas Söderlund, Wolfram Sang, linux-kernel
Cc: linux-renesas-soc, Phong Hoang
On 22/04/2021 14:34, Niklas Söderlund wrote:
> From: Phong Hoang <phong.hoang.wz@renesas.com>
>
> If CMT instance has at least two channels, one channel will be used
> as a clock source and another one used as a clock event device.
> In that case, IRQ is not requested for clock source channel so
> sh_cmt_clock_event_program_verify() might work incorrectly.
> Besides, when a channel is only used for clock source, don't need to
> re-set the next match_value since it should be maximum timeout as
> it still is.
>
> On the other hand, due to no IRQ, total_cycles is not counted up
> when reaches compare match time (timer counter resets to zero),
> so sh_cmt_clocksource_read() returns unexpected value.
> Therefore, use 64-bit clocksoure's mask for 32-bit or 16-bit variants
> will also lead to wrong delta calculation. Hence, this mask should
> correspond to timer counter width, and above function just returns
> the raw value of timer counter register.
I'm not getting the 'ch->cmt->num_channels == 1' change, can you explain?
> Fixes: bfa76bb12f23 ("clocksource: sh_cmt: Request IRQ for clock event device only")
> Fixes: 37e7742c55ba ("clocksource/drivers/sh_cmt: Fix clocksource width for 32-bit machines")
> Signed-off-by: Phong Hoang <phong.hoang.wz@renesas.com>
> Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
> ---
> drivers/clocksource/sh_cmt.c | 30 ++++++++++++++++++------------
> 1 file changed, 18 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c
> index c98f8851fd680454..cadd09ad1a0946b9 100644
> --- a/drivers/clocksource/sh_cmt.c
> +++ b/drivers/clocksource/sh_cmt.c
> @@ -578,7 +578,8 @@ static int sh_cmt_start(struct sh_cmt_channel *ch, unsigned long flag)
> ch->flags |= flag;
>
> /* setup timeout if no clockevent */
> - if ((flag == FLAG_CLOCKSOURCE) && (!(ch->flags & FLAG_CLOCKEVENT)))
> + if (ch->cmt->num_channels == 1 &&
> + flag == FLAG_CLOCKSOURCE && (!(ch->flags & FLAG_CLOCKEVENT)))
> __sh_cmt_set_next(ch, ch->max_match_value);
> out:
> raw_spin_unlock_irqrestore(&ch->lock, flags);
> @@ -620,20 +621,25 @@ static struct sh_cmt_channel *cs_to_sh_cmt(struct clocksource *cs)
> static u64 sh_cmt_clocksource_read(struct clocksource *cs)
> {
> struct sh_cmt_channel *ch = cs_to_sh_cmt(cs);
> - unsigned long flags;
> u32 has_wrapped;
> - u64 value;
> - u32 raw;
>
> - raw_spin_lock_irqsave(&ch->lock, flags);
> - value = ch->total_cycles;
> - raw = sh_cmt_get_counter(ch, &has_wrapped);
> + if (ch->cmt->num_channels == 1) {
> + unsigned long flags;
> + u64 value;
> + u32 raw;
>
> - if (unlikely(has_wrapped))
> - raw += ch->match_value + 1;
> - raw_spin_unlock_irqrestore(&ch->lock, flags);
> + raw_spin_lock_irqsave(&ch->lock, flags);
> + value = ch->total_cycles;
> + raw = sh_cmt_get_counter(ch, &has_wrapped);
>
> - return value + raw;
> + if (unlikely(has_wrapped))
> + raw += ch->match_value + 1;
> + raw_spin_unlock_irqrestore(&ch->lock, flags);
> +
> + return value + raw;
> + }
> +
> + return sh_cmt_get_counter(ch, &has_wrapped);
> }
>
> static int sh_cmt_clocksource_enable(struct clocksource *cs)
> @@ -696,7 +702,7 @@ static int sh_cmt_register_clocksource(struct sh_cmt_channel *ch,
> cs->disable = sh_cmt_clocksource_disable;
> cs->suspend = sh_cmt_clocksource_suspend;
> cs->resume = sh_cmt_clocksource_resume;
> - cs->mask = CLOCKSOURCE_MASK(sizeof(u64) * 8);
> + cs->mask = CLOCKSOURCE_MASK(ch->cmt->info->width);
> cs->flags = CLOCK_SOURCE_IS_CONTINUOUS;
>
> dev_info(&ch->cmt->pdev->dev, "ch%u: used as clock source\n",
>
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel
2021-06-10 12:45 ` Daniel Lezcano
@ 2021-06-11 7:56 ` Niklas Söderlund
2021-08-11 14:06 ` Wolfram Sang
0 siblings, 1 reply; 5+ messages in thread
From: Niklas Söderlund @ 2021-06-11 7:56 UTC (permalink / raw)
To: Daniel Lezcano; +Cc: Wolfram Sang, linux-kernel, linux-renesas-soc, Phong Hoang
Hi Daniel,
On 2021-06-10 14:45:05 +0200, Daniel Lezcano wrote:
> On 22/04/2021 14:34, Niklas Söderlund wrote:
> > From: Phong Hoang <phong.hoang.wz@renesas.com>
> >
> > If CMT instance has at least two channels, one channel will be used
> > as a clock source and another one used as a clock event device.
> > In that case, IRQ is not requested for clock source channel so
> > sh_cmt_clock_event_program_verify() might work incorrectly.
> > Besides, when a channel is only used for clock source, don't need to
> > re-set the next match_value since it should be maximum timeout as
> > it still is.
> >
> > On the other hand, due to no IRQ, total_cycles is not counted up
> > when reaches compare match time (timer counter resets to zero),
> > so sh_cmt_clocksource_read() returns unexpected value.
> > Therefore, use 64-bit clocksoure's mask for 32-bit or 16-bit variants
> > will also lead to wrong delta calculation. Hence, this mask should
> > correspond to timer counter width, and above function just returns
> > the raw value of timer counter register.
>
> I'm not getting the 'ch->cmt->num_channels == 1' change, can you explain?
My understanding is that if more then one channel is available the
channel used as clocksource is used without an interrupt. This was not
addressed in the patches listed as fixes. This patch fixes this multi
channel use-case while still retaining the old behavior for for the case
where only one channel is available (ch->cmt->num_channels ==
1).
>
> > Fixes: bfa76bb12f23 ("clocksource: sh_cmt: Request IRQ for clock event device only")
> > Fixes: 37e7742c55ba ("clocksource/drivers/sh_cmt: Fix clocksource width for 32-bit machines")
> > Signed-off-by: Phong Hoang <phong.hoang.wz@renesas.com>
> > Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
> > ---
> > drivers/clocksource/sh_cmt.c | 30 ++++++++++++++++++------------
> > 1 file changed, 18 insertions(+), 12 deletions(-)
> >
> > diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c
> > index c98f8851fd680454..cadd09ad1a0946b9 100644
> > --- a/drivers/clocksource/sh_cmt.c
> > +++ b/drivers/clocksource/sh_cmt.c
> > @@ -578,7 +578,8 @@ static int sh_cmt_start(struct sh_cmt_channel *ch, unsigned long flag)
> > ch->flags |= flag;
> >
> > /* setup timeout if no clockevent */
> > - if ((flag == FLAG_CLOCKSOURCE) && (!(ch->flags & FLAG_CLOCKEVENT)))
> > + if (ch->cmt->num_channels == 1 &&
> > + flag == FLAG_CLOCKSOURCE && (!(ch->flags & FLAG_CLOCKEVENT)))
> > __sh_cmt_set_next(ch, ch->max_match_value);
> > out:
> > raw_spin_unlock_irqrestore(&ch->lock, flags);
> > @@ -620,20 +621,25 @@ static struct sh_cmt_channel *cs_to_sh_cmt(struct clocksource *cs)
> > static u64 sh_cmt_clocksource_read(struct clocksource *cs)
> > {
> > struct sh_cmt_channel *ch = cs_to_sh_cmt(cs);
> > - unsigned long flags;
> > u32 has_wrapped;
> > - u64 value;
> > - u32 raw;
> >
> > - raw_spin_lock_irqsave(&ch->lock, flags);
> > - value = ch->total_cycles;
> > - raw = sh_cmt_get_counter(ch, &has_wrapped);
> > + if (ch->cmt->num_channels == 1) {
>
>
>
> > + unsigned long flags;
> > + u64 value;
> > + u32 raw;
> >
> > - if (unlikely(has_wrapped))
> > - raw += ch->match_value + 1;
> > - raw_spin_unlock_irqrestore(&ch->lock, flags);
> > + raw_spin_lock_irqsave(&ch->lock, flags);
> > + value = ch->total_cycles;
> > + raw = sh_cmt_get_counter(ch, &has_wrapped);
> >
> > - return value + raw;
> > + if (unlikely(has_wrapped))
> > + raw += ch->match_value + 1;
> > + raw_spin_unlock_irqrestore(&ch->lock, flags);
> > +
> > + return value + raw;
> > + }
> > +
> > + return sh_cmt_get_counter(ch, &has_wrapped);
> > }
> >
> > static int sh_cmt_clocksource_enable(struct clocksource *cs)
> > @@ -696,7 +702,7 @@ static int sh_cmt_register_clocksource(struct sh_cmt_channel *ch,
> > cs->disable = sh_cmt_clocksource_disable;
> > cs->suspend = sh_cmt_clocksource_suspend;
> > cs->resume = sh_cmt_clocksource_resume;
> > - cs->mask = CLOCKSOURCE_MASK(sizeof(u64) * 8);
> > + cs->mask = CLOCKSOURCE_MASK(ch->cmt->info->width);
> > cs->flags = CLOCK_SOURCE_IS_CONTINUOUS;
> >
> > dev_info(&ch->cmt->pdev->dev, "ch%u: used as clock source\n",
> >
>
>
> --
> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>
> Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog
--
Regards,
Niklas Söderlund
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel
2021-06-11 7:56 ` Niklas Söderlund
@ 2021-08-11 14:06 ` Wolfram Sang
0 siblings, 0 replies; 5+ messages in thread
From: Wolfram Sang @ 2021-08-11 14:06 UTC (permalink / raw)
To: Niklas Söderlund
Cc: Daniel Lezcano, linux-kernel, linux-renesas-soc, Phong Hoang
[-- Attachment #1: Type: text/plain, Size: 1476 bytes --]
Hi Daniel,
> > > If CMT instance has at least two channels, one channel will be used
> > > as a clock source and another one used as a clock event device.
> > > In that case, IRQ is not requested for clock source channel so
> > > sh_cmt_clock_event_program_verify() might work incorrectly.
> > > Besides, when a channel is only used for clock source, don't need to
> > > re-set the next match_value since it should be maximum timeout as
> > > it still is.
> > >
> > > On the other hand, due to no IRQ, total_cycles is not counted up
> > > when reaches compare match time (timer counter resets to zero),
> > > so sh_cmt_clocksource_read() returns unexpected value.
> > > Therefore, use 64-bit clocksoure's mask for 32-bit or 16-bit variants
> > > will also lead to wrong delta calculation. Hence, this mask should
> > > correspond to timer counter width, and above function just returns
> > > the raw value of timer counter register.
> >
> > I'm not getting the 'ch->cmt->num_channels == 1' change, can you explain?
>
> My understanding is that if more then one channel is available the
> channel used as clocksource is used without an interrupt. This was not
> addressed in the patches listed as fixes. This patch fixes this multi
> channel use-case while still retaining the old behavior for for the case
> where only one channel is available (ch->cmt->num_channels ==
> 1).
Did Niklas answer help you?
Happy hacking,
Wolfram
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* [tip: timers/core] clocksource/drivers/sh_cmt: Fix wrong setting if don't request IRQ for clock source channel
2021-04-22 12:34 [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel Niklas Söderlund
2021-06-10 12:45 ` Daniel Lezcano
@ 2021-08-26 16:25 ` tip-bot2 for Phong Hoang
1 sibling, 0 replies; 5+ messages in thread
From: tip-bot2 for Phong Hoang @ 2021-08-26 16:25 UTC (permalink / raw)
To: linux-tip-commits
Cc: Phong Hoang, niklas.soderlund+renesas, Daniel Lezcano, x86, linux-kernel
The following commit has been merged into the timers/core branch of tip:
Commit-ID: be83c3b6e7b8ff22f72827a613bf6f3aa5afadbb
Gitweb: https://git.kernel.org/tip/be83c3b6e7b8ff22f72827a613bf6f3aa5afadbb
Author: Phong Hoang <phong.hoang.wz@renesas.com>
AuthorDate: Thu, 22 Apr 2021 14:34:43 +02:00
Committer: Daniel Lezcano <daniel.lezcano@linaro.org>
CommitterDate: Fri, 13 Aug 2021 20:26:29 +02:00
clocksource/drivers/sh_cmt: Fix wrong setting if don't request IRQ for clock source channel
If CMT instance has at least two channels, one channel will be used
as a clock source and another one used as a clock event device.
In that case, IRQ is not requested for clock source channel so
sh_cmt_clock_event_program_verify() might work incorrectly.
Besides, when a channel is only used for clock source, don't need to
re-set the next match_value since it should be maximum timeout as
it still is.
On the other hand, due to no IRQ, total_cycles is not counted up
when reaches compare match time (timer counter resets to zero),
so sh_cmt_clocksource_read() returns unexpected value.
Therefore, use 64-bit clocksoure's mask for 32-bit or 16-bit variants
will also lead to wrong delta calculation. Hence, this mask should
correspond to timer counter width, and above function just returns
the raw value of timer counter register.
Fixes: bfa76bb12f23 ("clocksource: sh_cmt: Request IRQ for clock event device only")
Fixes: 37e7742c55ba ("clocksource/drivers/sh_cmt: Fix clocksource width for 32-bit machines")
Signed-off-by: Phong Hoang <phong.hoang.wz@renesas.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20210422123443.73334-1-niklas.soderlund+renesas@ragnatech.se
---
drivers/clocksource/sh_cmt.c | 30 ++++++++++++++++++------------
1 file changed, 18 insertions(+), 12 deletions(-)
diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c
index d7ed99f..dd0956a 100644
--- a/drivers/clocksource/sh_cmt.c
+++ b/drivers/clocksource/sh_cmt.c
@@ -579,7 +579,8 @@ static int sh_cmt_start(struct sh_cmt_channel *ch, unsigned long flag)
ch->flags |= flag;
/* setup timeout if no clockevent */
- if ((flag == FLAG_CLOCKSOURCE) && (!(ch->flags & FLAG_CLOCKEVENT)))
+ if (ch->cmt->num_channels == 1 &&
+ flag == FLAG_CLOCKSOURCE && (!(ch->flags & FLAG_CLOCKEVENT)))
__sh_cmt_set_next(ch, ch->max_match_value);
out:
raw_spin_unlock_irqrestore(&ch->lock, flags);
@@ -621,20 +622,25 @@ static struct sh_cmt_channel *cs_to_sh_cmt(struct clocksource *cs)
static u64 sh_cmt_clocksource_read(struct clocksource *cs)
{
struct sh_cmt_channel *ch = cs_to_sh_cmt(cs);
- unsigned long flags;
u32 has_wrapped;
- u64 value;
- u32 raw;
- raw_spin_lock_irqsave(&ch->lock, flags);
- value = ch->total_cycles;
- raw = sh_cmt_get_counter(ch, &has_wrapped);
+ if (ch->cmt->num_channels == 1) {
+ unsigned long flags;
+ u64 value;
+ u32 raw;
- if (unlikely(has_wrapped))
- raw += ch->match_value + 1;
- raw_spin_unlock_irqrestore(&ch->lock, flags);
+ raw_spin_lock_irqsave(&ch->lock, flags);
+ value = ch->total_cycles;
+ raw = sh_cmt_get_counter(ch, &has_wrapped);
+
+ if (unlikely(has_wrapped))
+ raw += ch->match_value + 1;
+ raw_spin_unlock_irqrestore(&ch->lock, flags);
+
+ return value + raw;
+ }
- return value + raw;
+ return sh_cmt_get_counter(ch, &has_wrapped);
}
static int sh_cmt_clocksource_enable(struct clocksource *cs)
@@ -697,7 +703,7 @@ static int sh_cmt_register_clocksource(struct sh_cmt_channel *ch,
cs->disable = sh_cmt_clocksource_disable;
cs->suspend = sh_cmt_clocksource_suspend;
cs->resume = sh_cmt_clocksource_resume;
- cs->mask = CLOCKSOURCE_MASK(sizeof(u64) * 8);
+ cs->mask = CLOCKSOURCE_MASK(ch->cmt->info->width);
cs->flags = CLOCK_SOURCE_IS_CONTINUOUS;
dev_info(&ch->cmt->pdev->dev, "ch%u: used as clock source\n",
^ permalink raw reply related [flat|nested] 5+ messages in thread
end of thread, other threads:[~2021-08-26 16:25 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-04-22 12:34 [PATCH] clocksource: sh_cmt: Fix wrong setting if don't request IRQ for clock source channel Niklas Söderlund
2021-06-10 12:45 ` Daniel Lezcano
2021-06-11 7:56 ` Niklas Söderlund
2021-08-11 14:06 ` Wolfram Sang
2021-08-26 16:25 ` [tip: timers/core] clocksource/drivers/sh_cmt: " tip-bot2 for Phong Hoang
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.